Hi,
managed to create sample with chromadb, here are most important parts:

Fix for huggingface embeddings and chromadb latest version

from chromadb.utils import embedding_functions
class MyEmbeddingFunction(EmbeddingFunction[Documents]):
def call(self, input: Documents) -> Embeddings:
sentence_transformer_ef = embedding_functions.SentenceTransformerEmbeddingFunction(model_name="intfloat/multilingual-e5-large")
embeddings = sentence_transformer_ef(input)
return embeddings
custom = MyEmbeddingFunction()

Initialize the chromadb directory, and client.

chroma_client = chromadb.PersistentClient(path="./chroma/RAG_DB")
chroma_collection = chroma_client.get_or_create_collection(name=f"ProductsRag", embedding_function=custom)

I can add products to DB and as it lasts some time it looks like it is using correct sentence transformer:
chroma_collection.add(
documents=products_list,
ids=ids
)

and then I can successfully query the cromadb:
queryStr = "Some question to ask"
results = chroma_collection.query(
query_texts=queryStr,
n_results=10,
include=['documents', 'distances']
)
print(len(results["documents"][0]))
i = 0
for value in results["documents"][0]:
i = i + 1
print(f"DOC {i}:")
print(value)